Psychiatric disorders are an increasingly serious health problem in developed countries. Although past studies on families, twins, and adoptees have provided convincing evidence that genetics influence psychiatric and substance use disorders, such studies have identified only a moderate proportion of genetically-based psychiatric and comorbid disorders. The underlying etiology is thus likely explained by a combination or interaction of genetic and environmental risk factors. To date, few or no studies have examined this combination in representative samples as large as those in the proposed study. Objectives: Contribute to disentangling genetic and environmental influences on psychiatric disorders to guide future research and lend insight into new underlying mechanisms, in order to help reduce the burden of psychiatric disorders. Specific aims: Building on partnerships among the Lund University, Stanford University, and the Virginia Commonwealth University, this cross-institutional study will be based on cross-sectional and longitudinal data and will: 1) Estimate the risk of psychiatric disorders in first-degree relatives (who share both genetics and family environment) compared with second-degree relatives (who share genetics but not family environment) of at least one affected proband, further modeled in dizygotic and monozygotic twins using Structural Equation Modeling, 2) Compare the risk of psychiatric and substance use disorders in non-genetically related individuals (adoptees) who share adult and childhood family environments with risk in genetically-related people, and 3) Examine whether neighborhood-level factors exert a differential effect on risk of psychiatric disorders in genetically susceptible individuals, defined as first-degree or second-degree relatives of an affected proband. Design/methods: The relative contributions of genetic and environmental factors to the occurrence of psychiatric and substance use disorders and the differential risk of psychiatric disorders will be estimated by Structural Equation Modeling and Multilevel and General Linear Models when appropriate. All persons in Sweden have a personal identification number that will be used to obtain and link the following data for the total Swedish population: annual census data; cause of death records; prescription medicine records; multigenerational family register records; neighborhood-level social and physical environmental records, including geocoded goods and services, and on a large subsample, inpatient and outpatient hospital and clinic records that include all psychiatric care records. Diagnoses of psychiatric and substance use disorders are available from 1962 (inpatient) and 2001 (outpatient), and individual- and neighborhood-level factors from 1960. Data are available on 11.2 million people, living and deceased, among whom there are 3.2 million nuclear families, 7.5 million offspring, 172,000 twins, and 136,000 adoptees. The linked data will form a unique, extensive, and comprehensive new Mental Risk Database. Group-level data will be accessible to researchers in the United States via a link from the NIMH Web page under conditions of technical assistance.